Use of color information for structured-light 3D shape measurement of objects with shiny surfaces

Objects with shiny surfaces cannot be directly measured using the conventional structured-light method. To cope with this problem, in this paper, we propose a novel method for removing the saturated components in an image. First, the specular pixels in the image are identified using a highlight extraction algorithm. Then, based on the reflection component separation (RCS) theory, the diffuse and specular components of these specular pixels are separated. For objects with shiny surfaces, use of the RCS approach destroys the color information of highlighted pixels with a large specular reflection component. As such, finally, the color information of the highlighted pixels is recovered using an image repair process. Experimental results indicate that 95% of the highlights in the images were eliminated. The highlight caused by the strong reflecting surface can be well suppressed. This proposed method effectively overcomes the interference of reflected light and provides a feasible solution to the problems associated with the structured-light measurement of objects with smooth surfaces.